gtorbsim does not implement the rocking profile correctly

[joeastro]

When using gtorbsim to create a predicted FT2 file using a Science Timeline provided by TAKO, the spacecraft attitude and pointing data do not match the values shown in TAKO during timeline creation. Examination of the ROCK_ANGLE column values in particular indicated that the rocking profile repeat period(s) used by gtorbsim was somehow different than the repeat period(s) provided in the science timeline. This behavior first manifested in Mission Week 805 and has become more pronounced since.

The cause of the issue has been traced to the file read_ephem.cxx, which "contains functions to read/calculate ephemerides and the attitude calculation." Specifically, the variable "difincr" is mishandled in the current implementation. When gtorbsim is run the function InterpProfile is called, which recasts the 17 point profile defined by the mission planners so that the timings of the 17 points are multiples of 60 (because the crated FT2 file has one line for each minute, starting at the top of the minute). The 17th interpolated point is thus usually incorrect, and so the the "difincr" variable is used to track this difference and make a correction down-stream.

However, as currently written the value of "difincr" can only be positive and the resulting correction can only be positive as well. This issue was not noticed until now because between launch and MW 805 the needed correction was always positive, but after the the repeat period fell below 5700s (due to decreasing orbit altitude), a negative correction has been needed.

I have made the necessary modifications to the code in read_ephem.cxx and tested them on the timelines of several mission weeks, both before and after MW 805. Results are now as expected, so I will initiate a Pull Request soon so that these changes can be incorporated into the tools.